Menu
▾
▴
ipodlinux-devel
You can subscribe to this list here.
| 2003 |
Jan
|
Feb
(9) |
Mar
(21) |
Apr
(12) |
May
(11) |
Jun
(6) |
Jul
(3) |
Aug
|
Sep
(10) |
Oct
(20) |
Nov
(32) |
Dec
(41) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2004 |
Jan
(48) |
Feb
(27) |
Mar
(120) |
Apr
(69) |
May
(15) |
Jun
(1) |
Jul
(6) |
Aug
(1) |
Sep
(1) |
Oct
(6) |
Nov
(2) |
Dec
(4) |
| 2005 |
Jan
|
Feb
(10) |
Mar
(1) |
Apr
(1) |
May
|
Jun
(3) |
Jul
|
Aug
(2) |
Sep
(2) |
Oct
|
Nov
(1) |
Dec
(1) |
| 2006 |
Jan
(3) |
Feb
(4) |
Mar
(9) |
Apr
(2) |
May
|
Jun
|
Jul
(1) |
Aug
(1) |
Sep
|
Oct
|
Nov
(2) |
Dec
(2) |
| 2007 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(5) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2015 |
Jan
|
Feb
|
Mar
|
Apr
(1) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Bernard L. <le...@bo...> - 2003-03-03 11:46:34
|
Hi David, What are the main differences between HFS and HFS+? Is HFS supported as a Linux root filesystem? If not, we could just use the ROMFS filesystem to include a root filesystem with the kernel (e.g. it would be stored in the boot partition). The HFS partition would then just be used to store the application data (music & configuration). cheers, bern. On Fri, 2003-02-28 at 17:54, David Carne wrote: > I've been experimenting around with UFS for linux so iTunes will > automatically update for mac users and I have some good news and bad > news: > > Good: > -UFS works on the iPod > > Bad: > -UFS causes a kernel panic on the mac. > > It works for a while, but I suspect iTunes may be doing something funky > with the UFS FAT.... Mistaking it for an HFS one and putting some > values in that shouldn't be there. Anyways, after the first iTunes > update, I get a kernel panic whenever the iPod is connected... So it > rules out that line of exploration. > > I'll look into just using HFS now... see whether that works ok. > Later, > --David Carne > > > > ------------------------------------------------------- > This sf.net email is sponsored by:ThinkGeek > Welcome to geek heaven. > http://thinkgeek.com/sf > _______________________________________________ > iPodlinux-devel mailing list > iPo...@li... > https://lists.sourceforge.net/lists/listinfo/ipodlinux-devel > |
|
From: Bernard L. <le...@bo...> - 2003-03-03 11:33:52
|
Hi Chandan, A couple of comments. There is a newer version of the Intel libraries that may also provide some improvement. Last time I looked there was a 3.0 download but I couldnt install the thing as I dont have a RPM system (or the patience to stuff about with them). These libraries also have JPEG processing which could be quite cool. The other audio decoder option is madplay. I was using that originally but not getting very good performance. Andre (see the CC list) sent me some ARM performance patches which should help. Regarding the playlists etc have you looked at any of the existing Linux programs for processing the current iPod databases? I imagine a lot of Mac users would be quite happy to keep using iTunes to manage their stuff... cheers, bern. On Tue, 2003-02-25 at 18:23, Chandan Kudige [home] wrote: > I have been fiddling around with the intel mp3example code and with the > latest changes fron Bernard, I can play most mp3 songs with as little as > one or two skips. Some songs, however, have more skips but thats rare. > > I have started working on a simple shell to navigate filesystem and play > mp3 files. It works quite decently and it can also log the songs played > on to the hard-disk for future reference. Writing to the disk seems to > be slow the first time, but after that there is not much perceptible > delay. By creating directory heirarchy based on the album and artist > tags and symbolic links to actual files, we can simulate the apple's > firmware for navigating the songs. Playlists can be created by creating > separate directories with symbolic links. > > There seems to be enough bandwidth for one screen update and handling > couple of key strokes between each frame being decoded. This should be > decent enough for most purposes. > > I also compiled the ffmpeg (http://ffmpeg.sf.net) decoder which mostly > integer based (except for initial header parsing), but the performance > is quite bad (80% real-time). So for now we have to stick with the > intel's code which cannot be re-distributed. > > At this point I am interested in hearing feature requests for the > firmware. I have a couple in mind: > > 1. Filesystem navigation > 2. Playlist queuing > 3. Tagging songs at runtime > 4. Preserving runtime song information : order in which songs were > played, play count for each song, number of minutes each song was played > etc. can be very useful in managing and pruning huge song collection. > 5. Resume across power-down > 6. ID3 support > > I have also managed to display digital images (although at a much lower > resolution) with some preprocessing with gimp (converting to 4 level > grayscale). We can add this as an addition feature (photo album). > > Of course, no handheld is complete without atleast a few games (sokoban, > tetris, bricks etc?) > > More advanced options would be: > - bookmarking song positions > - Firewire detach/attach without reboot (depends on the firewire driver > code) > - Automatic volume normalisation etc. > > Cheers, > Chandan > > > |
|
From: David C. <bus...@ho...> - 2003-02-28 16:59:08
|
I've been experimenting around with UFS for linux so iTunes will automatically update for mac users and I have some good news and bad news: Good: -UFS works on the iPod Bad: -UFS causes a kernel panic on the mac. It works for a while, but I suspect iTunes may be doing something funky with the UFS FAT.... Mistaking it for an HFS one and putting some values in that shouldn't be there. Anyways, after the first iTunes update, I get a kernel panic whenever the iPod is connected... So it rules out that line of exploration. I'll look into just using HFS now... see whether that works ok. Later, --David Carne |
|
From: Dan C. <da...@da...> - 2003-02-25 20:34:12
|
Hi, A friend of mine wanted to try running Linux on his iPod, but he didn't have a PC to test it on. The patch tool didn't work on Darwin, and I found out that these issues were related to the big-endian PPC processor. Attached below is a version of the patch tool which works on Darwin. It would probably work on any other big-endian processors and operating system as well. - Dan Christiansen |
|
From: Chandan K. [home] <ch...@to...> - 2003-02-25 17:23:23
|
I have been fiddling around with the intel mp3example code and with the latest changes fron Bernard, I can play most mp3 songs with as little as one or two skips. Some songs, however, have more skips but thats rare. I have started working on a simple shell to navigate filesystem and play mp3 files. It works quite decently and it can also log the songs played on to the hard-disk for future reference. Writing to the disk seems to be slow the first time, but after that there is not much perceptible delay. By creating directory heirarchy based on the album and artist tags and symbolic links to actual files, we can simulate the apple's firmware for navigating the songs. Playlists can be created by creating separate directories with symbolic links. There seems to be enough bandwidth for one screen update and handling couple of key strokes between each frame being decoded. This should be decent enough for most purposes. I also compiled the ffmpeg (http://ffmpeg.sf.net) decoder which mostly integer based (except for initial header parsing), but the performance is quite bad (80% real-time). So for now we have to stick with the intel's code which cannot be re-distributed. At this point I am interested in hearing feature requests for the firmware. I have a couple in mind: 1. Filesystem navigation 2. Playlist queuing 3. Tagging songs at runtime 4. Preserving runtime song information : order in which songs were played, play count for each song, number of minutes each song was played etc. can be very useful in managing and pruning huge song collection. 5. Resume across power-down 6. ID3 support I have also managed to display digital images (although at a much lower resolution) with some preprocessing with gimp (converting to 4 level grayscale). We can add this as an addition feature (photo album). Of course, no handheld is complete without atleast a few games (sokoban, tetris, bricks etc?) More advanced options would be: - bookmarking song positions - Firewire detach/attach without reboot (depends on the firewire driver code) - Automatic volume normalisation etc. Cheers, Chandan |
|
From: Bernard L. <le...@bo...> - 2003-02-18 08:41:28
|
On Tue, 2003-02-18 at 00:55, Chandan Kudige [home] wrote: > > That pretty much covers whats gone so far. For the next short while I > > plan on continuing with the firewire controller. If anyone has > > suggestions on how integrate things like the remote control or power > > management I'd be happy to hear them. > Here is just a clue towards the power management. I made a small change to the > reboot_to_diskmode(): > > > /* do special startup? */ > > outl(0, 0x40017f10); > > > > /* reset */ > > outl(inl(0xcf005030) | 0xb, 0xcf005030); > > And found that this caused the ipod to 'freeze'. It didnt respond to any key strokes. The > LCD display froze at what was just displayed and CPU froze too. The only way to get out of > this is the '2 finger salute' (menu + play for a few seconds). Not sure, but the this > may be part of the power management. Hi Chandan, The write to 0x40017f10 needs to be a 1 to indicate that we want the booloader to start a particular image rather than just boot the default. As for the 0xcf005030 regster 0x400 resets the LCD 0x2 resets the coprocessor 0x4 resets the system 0xf3e0 probably resets a whole bunch of things I think the 0xcf005034 register is also a reset register but I've only seen 0x7 and 0x2 written there. Regarding the power management, I'm interested in finding out how linux drivers can contribute to reducing the overall power usage. This would be key to a successful firwmare! cheers, bern. |
|
From: Chandan K. [home] <ch...@to...> - 2003-02-18 02:31:01
|
Bernard, > Since the initial patch I've committed two changes. The most > interesting one is some fix-ups to the frame buffer (LCD) driver. > Firstly the annoying flashes on startup are gone but even better the > performance has increased dramatically (so much so that now with the > intel MP3 player I'm no longer getting any play glitches!!). The second The LCD performance is really fast, and the performance of the intel player is quite nice. But I see some tiny glitches occasionally. But importantly key-presses and scrolling the screen no longer breaks the song play which is a good news. Couple of questions: Whats the significance of the hardcoded addresses used as the buffer between the main CPU and the co-processor ? It is quite small in size, and I was wondering if making this bigger might help. I did try using a malloced buffer but the performance did not improve. Have you considered pushing the whole mp3 decoding onto the co-processor ? I noticed that if I load a chunk of mp3 into ram and feed it, there are no glitches whatsoever. This can be easily done if we offload the decoding onto the coproc. -Chandan |
|
From: Chandan K. [home] <ch...@to...> - 2003-02-17 23:56:09
|
> Since the initial patch I've committed two changes. The most > interesting one is some fix-ups to the frame buffer (LCD) driver. > Firstly the annoying flashes on startup are gone but even better the > performance has increased dramatically (so much so that now with the > intel MP3 player I'm no longer getting any play glitches!!). The second This is great. I will immediately give it a spin ! > That pretty much covers whats gone so far. For the next short while I > plan on continuing with the firewire controller. If anyone has > suggestions on how integrate things like the remote control or power > management I'd be happy to hear them.Here is just a clue towards the power management. I made a small change to the reboot_to_diskmode(): > /* do special startup? */ > outl(0, 0x40017f10); > > /* reset */ > outl(inl(0xcf005030) | 0xb, 0xcf005030); And found that this caused the ipod to 'freeze'. It didnt respond to any key strokes. The LCD display froze at what was just displayed and CPU froze too. The only way to get out of this is the '2 finger salute' (menu + play for a few seconds). Not sure, but the this may be part of the power management. -Chandan |
|
From: Bernard L. <le...@bo...> - 2003-02-17 18:47:29
|
Hi all, Its been a busy couple of weeks since the site first went up, so I thought I'd write a short update to let you know whats new. Firstly I've moved my CVS repository onto sourceforge. This is now the best way to keep up-to-date with the latest changes. I don't plan on creating complete patches that often but I'll try to keep my changes in CVS. I've added some basic information on how to build with the CVS version on the "build instructions" page. For those that are interested in helping out with the active development please let me know so we can sort out the details. Since the initial patch I've committed two changes. The most interesting one is some fix-ups to the frame buffer (LCD) driver. Firstly the annoying flashes on startup are gone but even better the performance has increased dramatically (so much so that now with the intel MP3 player I'm no longer getting any play glitches!!). The second patch adds support for the scroll wheel (thanks to Chandan for sorting out the protocol!) and hold switch. I've also done some preliminary work (not in CVS yet) for sorting out the firewire controller. Sofar I can talk to the chip and see interrupts occuring so that at least is a start! That pretty much covers whats gone so far. For the next short while I plan on continuing with the firewire controller. If anyone has suggestions on how integrate things like the remote control or power management I'd be happy to hear them. cheers, bern. |
|
From: Chandan K. [home] <ch...@to...> - 2003-02-06 03:56:11
|
Hi, I managed to boot the linux kernel (built following the instructions on the web-page) and I am trying to setup the root filesystem. I dont see the rootfs download, where can I get it from ? I compiled a small 'init' program to open the console and print a hello message, but I find that /dev is empty, no /dev/console. I also get a message "Unable to open the initial console" before loading the init. But I verified that my init is indeed getting loaded. How to mount /dev ? Should I enable devfs ? Any help appreciated. BTW, I noticed that the kernel available for download expects root to be on /dev/hda3. Is this intentional ? My root partition seems to be at /dev/hda2. Thanks, Chandan |
|
From: Chandan K. [home] <ch...@to...> - 2003-02-06 03:56:11
|
Hi, I managed to boot the linux kernel (built following the instructions on the web-page) and I am trying to setup the root filesystem. I dont see the rootfs download, where can I get it from ? I compiled a small 'init' program to open the console and print a hello message, but I find that /dev is empty, no /dev/console. I also get a message "Unable to open the initial console" before loading the init. But I verified that my init is indeed getting loaded. How to mount /dev ? Should I enable devfs ? Any help appreciated. BTW, I noticed that the kernel available for download expects root to be on /dev/hda3. Is this intentional ? My root partition seems to be at /dev/hda2. Thanks, Chandan |
5 messages has been excluded from this view by a project administrator.
